A number of processing systems exist for use during the manufacture of semiconductor wafers, magnetic memory disks, optical devices, and other delicate workpieces. For example, chemical mechanical polishing (CMP) machines are utilized to planarize the surfaces of semiconductor wafers, and wafer cleaning machines are utilized to clean and rinse the surfaces of such semiconductor wafers following the planarization process. Processing speed is an important factor in the semiconductor fabrication industry because increased product throughput relates directly to increased profitability. Accordingly, CMP machines, cleaning systems, and other processing devices are being developed in a manner that increases efficiency while maintaining reliability and robustness. Indeed, some semiconductor processing systems now utilize several combined subsystems in lieu of a number of individual distinct machines.
Semiconductor processing systems that combine several functions (e.g., polishing, cleaning, and rinsing) typically require intricate electronic control systems. Consequently, the efficiency and reliability of any given processing system may be directly related to the design and architecture of the control system. Furthermore, the cost of installation, testing, troubleshooting, and maintenance of the processing system may also be directly related to the complexity of the control system and how the control system interacts with the various functional components and subsystems of the processing system.
A prior art control system for a multiple function wafer processing system may employ a plurality of control modules concentrated in one central location, where each of the control modules are dependent upon a master controller. Such an architecture does not lend itself to independent testing and troubleshooting of the individual subsystems. In other words, such control systems require the entire processing system to be assembled and tested as an integral whole. Prior art control systems may employ individual control modules that are considered as "best in class" for their particular functions. Unfortunately, the "best in class" modules may be manufactured by different companies, which may present compatibility problems and cause the overall control system to be difficult to support and maintain. In addition, the use of "best in class" components may not necessarily result in the best combined control system.
Prior art control systems may not be adapted to independently control the different subsystems in a manner that enables continuous batch-specific processing with different polishing or cleaning parameters. Such prior art systems may be limited to the use of one processing recipe per run of wafers. Furthermore, such prior art control systems may control the operation of the processing system via serial instructions; the individual subsystems may be configured in a daisy chain arrangement. The use of serial instructions may lead to an unacceptable amount of wait time between processing stations and, consequently, an undesirably low processing throughput.
A prior art control system program for a semiconductor processing system may be integrally designed along with the application software associated with the processing system. With such an arrangement, changes to the control system hardware or software (e.g., type of processor, memory capacity, operating system, etc.) may require substantial modifications to the control system operating software. Similarly, any changes to the application-specific code associated with the processing system may require corresponding changes to the control system operating software.